Southern Cross University

ePublications@SCU
23rd Australasian Conference on the Mechanics of Structures and Materials

2014

The effect of prestress force magnitude on the

natural bending frequencies of prestressed concrete
structures
D Noble
Trinity College Dublin

M Nogal
Trinity College Dublin

A J. O'Connor
Trinity College Dublin

V Pakrashi
University College Cork, Ireland

Publication details
Noble, D, Nogal, M, O'Connor, AJ & Pakrashi, V 2014, 'The effect of prestress force magnitude on the natural bending frequencies of
prestressed concrete structures', in ST Smith (ed.), 23rd Australasian Conference on the Mechanics of Structures and Materials
(ACMSM23), vol. I, Byron Bay, NSW, 9-12 December, Southern Cross University, Lismore, NSW, pp. 333-338. ISBN:
9780994152008.

ePublications@SCU is an electronic repository administered by Southern Cross University Library. Its goal is to capture and preserve the intellectual
output of Southern Cross University authors and researchers, and to increase visibility and impact through open access to researchers around the
world. For further information please contact epubs@scu.edu.au.
 23rd Australasian Conference on the Mechanics of Structures and Materials (ACMSM23)
Byron Bay, Australia, 9-12 December 2014, S.T. Smith (Ed.)

THE EFFECT OF PRESTRESS FORCE MAGNITUDE ON THE NATURAL

BENDING FREQUENCIES OF PRESTRESSED CONCRETE STRUCTURES

D. Noble*
Dept. of Civil, Structural & Environmental Engineering, Museum Building,
Trinity College Dublin, College Green, Dublin 2, Ireland. nobleda@tcd.ie (Corresponding Author)

M. Nogal
Dept. of Civil, Structural & Environmental Engineering, Museum Building,
Trinity College Dublin, College Green, Dublin 2, Ireland. nogalm@tcd.ie

A.J. O’Connor
Dept. of Civil, Structural & Environmental Engineering, Museum Building,
Trinity College Dublin, College Green, Dublin 2, Ireland. oconnoaj@tcd.ie

V. Pakrashi
Dynamical Systems and Risk Laboratory, Civil & Environmental Engineering,
School of Engineering, University College Cork, Ireland. V.Pakrashi@ucc.ie

ABSTRACT

The effect of prestress force magnitude (N) and eccentricity (e) on natural frequencies (NFs, ω n) of
prestressed concrete (PSC) structures has been widely debated. No agreement currently exists as to
how prestress force magnitude and eccentricity affect natural frequencies of PSC structures. Some
reports state that the NFs decrease with increasing N (“compression softening” effect); others suggest
that NFs increase with increasing N. Finally, some claim that the NFs are unaffected by N. This paper
describes and analyses impact hammer testing conducted on a post-tensioned concrete beam at various
levels of post-tensioning force. The aim of the research is to investigate how the NFs of PSC structures
change with prestress force magnitude. Experimental modal analysis has been conducted and the
experimental natural frequencies were identified by employing the Fast Fourier Transform (FFT). This
was repeated for varying values of post-tensioning force at different impact locations. The applications
of the results are widespread. The Eurocode 2 (EC2) equation for prestress loss as a function of time is
well established. If the relationship between prestress force and NF for PSC structures can be
established, then the variation of NF over the design life of a PSC structure due to prestress loss can be
estimated. This problem is particularly pertinent in the field of post-tensioned concrete wind turbine
towers and in the field of pre and post-tensioned concrete bridge girders.

KEYWORDS

Dynamics, fast fourier transform, impact hammer testing, natural frequency, post-tensioned concrete,
prestress force, prestressed concrete, signal processing.

INTRODUCTION

The effect that the applied prestressing force has on the dynamic behaviour of PSC structures is a
widely debated topic (Quilligan et al., 2012). There are currently three distinct arguments found in the
literature;

This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit
http://creativecommons.org/licenses/by/4.0/
333
 1. The natural vibration frequency (ωn) decreases with increasing prestress force magnitude (N).
This is known as the “compression softening” effect and is based on classical Euler-Bernoulli
theory of an externally axially loaded homogeneous beam (Miyamoto et al., 2000; Tse et al.,
1978).
2. ωn is unaffected by N. This is based on a non-linear kinematic model which concludes that the
final equation of motion for a vibrating beam is independent of prestress force magnitude
(Hamed & Frostig, 2006).
3. ωn increases with increasing N. This has found to be the case in numerous empirical studies
(Hop, 1991; Saiidi et al., 1994) however, a satisfactory mathematical model has yet to be
formulated, despite some attempts (Kim et al., 2004).

The importance of this research is widespread. The effect of prestress force magnitude on NFs of PSC
structures has many implications, specifically in the PSC bridge industry and for pre-cast, post-
tensioned concrete wind turbine towers. Prestress force decreases over time due to concrete creep,
steel relaxation, anchorage pull in and other factors. It should be possible to monitor or estimate
changes in the NFs of PSC structures over the course of their design life to ensure their safety and
serviceability. As a result, prediction of change in NFs of prestressed concrete structures over time is
of great importance.

EXPERIMENTAL SET-UP

Impact Hammer Testing

Dynamic impact hammer testing was conducted on a post-tensioned concrete beam. The properties of
the concrete beam tested are outlined in Table 1.

Table 1. Properties of concrete beam tested

Property Value Unit
Height, ℎ 200 mm
Breadth, 𝑏 150 mm
Young’s Modulus, 𝐸 26.88 GPa
Characteristic strength, 𝑓𝑐𝑘 30 MPa
Density, 𝜌𝑐 23.55 kN/m3
Span, 𝑙 2.0 m

Strength testing was also carried out on standard concrete cube specimens and an average of 42MPa
was obtained, which is consistent with a characteristic strength of 30MPa (Table 1). Minimum
reinforcement was provided in accordance with Eurocode 2. Cover to all reinforcement was specified
as 25mm. 2No. H8 hanger bars were specified as top reinforcement, while 2No. H12 were specified as
main (bottom) reinforcement. Shear links (H8) were provided at 200mm centres. Extra shear
reinforcement was provided in the beam ends to prevent bursting failure due to localised stresses
during post-tensioning, in accordance with the CIRIA method. 4No. H10 U-bars were provided at the
beam ends for anchorage and continuity of reinforcement. A 20mm straight-profiled plastic duct was
cast through the centroid of the concrete section, through which the post-tensioning strand was
threaded. The estimated first natural frequency, taking into account the geometric effects of
reinforcement with Es = 205GPa, and post-tensioning strand on the second moment of area of the
section was estimated to be 78.05Hz.

The experiment was set-up as shown in Figure 1. A 15.7mm Freyssinet prestressing strand, with a
yield strength, fy=1880MPa was threaded through the straight-profiled duct. A 20mm thick steel plate
was placed against the rectangular beam face at each end. 300 ton hydraulic loading jacks were then
placed at each end of the beam. The jacks were connected to load cells, which measured the magnitude
of the post-tensioning force induced in the beam. The strand was secured using collets on either end of
the load cells. An impact hammer rig was assembled using a rope and pulley system.

ACMSM23 2014 334

 Figure 1. Experimental set-up 1
Impact hammer testing was conducted at three locations along the length of the beam span, as outlined
in Figure 2, labelled L1-L3. The beam was struck 10 times at each location using an impact hammer
rig assembled in the lab with a rope and pulley system. This helped ensure repeatability of the
experiment. Strain gauges were affixed at the three impact locations (ε1- ε3) in order to obtain the
mode shapes of vibration. The accelerometer (A1) was strategically placed at a distance of 800mm
from the support, in order to identify all of the first three modes of vibration. Placement at midspan
would eliminate the opportunity to obtain the second mode of vibration as it is a nodal point for the
second mode. A fourth strain gauge (εa1) was placed in the axial direction, close to midspan, in order
to compare the axial strain data with the prestress load data obtained from the load cells.

Figure 2. Instrumentation set-up on concrete beam

SIGNAL PROCESSING

The raw acceleration-time data obtained from a StrainSmart 6000 data acquisition system was
analysed using a Fast Fourier Transform (FFT) in Matlab and the peaks of the absolute values of the
transform were located in the frequency domain. It was found that the unprocessed signals were too
noisy to correctly identify the correct modes of vibration. Noise was then removed from the signal
through a signal processing technique developed by the authors. The signal was first “centred”, using
a moving average filter. “Zero drift” was then removed from the signal, by shifting the signal so that it
oscillated about zero. Electrical noise at 50Hz and all harmonics were removed using a “notch” filter.
Then a 40Hz, 6th order high-pass butterworth filter was applied to the signals, to remove all low
frequency components. The fundamental frequency was expected to be approximately 78Hz. Finally,
the accelerometer data was then smoothed in the time domain to obtain the final processed input signal.
Following this, the FFT was performed on the filtered signal to obtain the frequency response function
(FRF). This FRF was then smoothed and the peaks were identified. Figure 3(a) shows the raw
(unprocessed) signal in blue, in both the time and frequency domains, and also the filtered (processed)
signal in red. As shown in Figure 3(a), the peaks are more readily identified using the filtered signal
(red) rather than the raw data. Figure 3(b) shows the filtered signal in both the time and the frequency

ACMSM23 2014 335

domain. The green line shows the signal after it has been smoothed in the frequency domain, and the
peaks have been identified (yellow data points).

(a) Signal Processing (b) FFT and peak identification

Figure 3. Signal processing, FFT and peak identification

RESULTS AND DISCUSSIONS

Figure 4(a) shows the results of the experimental modal analysis of the accelerometer impact hammer
data. 30 data points were collected at each prestress load level. Assuming normal distribution of the
data points at each load level, the standard deviation of the data ranges between 2.49 and 5.26 Hz. The
mean frequencies at each post-tension load level were joined (dotted blue line). No clear increasing or
decreasing trend in natural frequency with varying post-tensioning force was observed. A linear
regression analysis was carried out to determine if there is an increasing or decreasing trend in the
frequency with changing post-tension force magnitude, Figure 4(b).

(a) Observed Data (b) Least square regression of data

Figure 4. Regression analysis of Frequency vs. Post tensioning force

The regression analysis indicates that there is no statistically significant change in fundamental
frequency with increasing post-tensioning force (Table 2). The linear regression equation is given as;

𝜔1 = 68.6585 + 0.0033𝑁 (1)

ACMSM23 2014 336

𝜔1 is the fundamental frequency in Hz (response variable) and 𝑁 is the post-tensioning force in kN
(predictor variable). The 𝑅 2 value is 2.44×10-3, indicating no correlation between post-tensioning force
and fundamental frequency. The 𝑅 2 parameter measures the goodness of fit of the data to the
regression line. A 𝑅 2 value close to 1 represents a good fit, whereas a 𝑅 2 statistic close to 0 indicates
little or no relationship between the regression variables. Table 2 shows 95% confidence intervals on
the value of the linear regression parameters, intercept and slope. As indicated, we can strongly reject
the null hypothesis that the long run value of the intercept is zero, and conclude that it lies between
67.2023Hz and 70.1147Hz. We do not reject the null hypothesis that the long run value of the slope
parameter is equal to 0, and subsequently conclude that the fundamental frequency is independent of
prestress force magnitude.

Table 2. Regression Parameters

Regression Value Std. t-value tc p 95% CI
Parameter Error
Intercept 68.6585 0.6437 106.6604 2.2622 0.0000 (67.2023,70.1147)
Slope 0.0033 0.0054 0.6124 2.2622 0.2777 (-0.0090,0.0156)

Following a test on the normality of the data, as shown in Figure 5(a), it was shown that the data does
not follow a normal distribution. The p-value is < 0.0005, and the assumption of normality of data
must be rejected under a significance level of α=0.05. This brings the validity of the regression
analysis into question, as one of the main prerequisites for conducting these analyses is the assumption
of data normality. It is postulated that the lack of normality is due to the beam being tested over a two
day period, and subsequently subject to temperature fluctuations that may affect the modal properties
of the test specimen.

(a) Normality test (b) Standard residuals vs. fitted values

Figure 5. Residual analysis

Figure 5(b) shows the relationship between the standardised residuals and the fitted data. There is no
reason, from this data to conclude that the long run standard deviation of the data is different for each
fitted value. Furthermore, only two outliers have been identified (standardised residuals > +3 or < -3),
at a significance level of α=0.05. For a data set containing 330 data points, this is not considered to be
unusual.

CONCLUSIONS

From the experimental modal analysis conducted on the post-tensioned concrete beam with zero
eccentricity, and the subsequent regression analysis, it can be concluded, from a statistical standpoint
that, for a straight profiled prestress strand, no change in natural bending frequency was observed, and
that the prediction of natural frequency is independent of prestress force magnitude over the given

ACMSM23 2014 337

range of prestress force magnitudes. This concurs with the mathematical model outlined by Hamed
and Frostig (2006). Saiidi et al. (1994) report that the natural frequency increased with increasing post-
tensioning force, when they conducted similar studies, however, they did report the presence of
cracking at midspan in their tested beam specimen, and speculated that the increase may be attributed
to crack closure as the post-tensioning force increased. In this case, it was ensured that cracking did
not occur. Furthermore, their study did not include a statistical perspective. It should be noted however,
that statistical acceptance or rejection of the null hypotheses does not necessarily mean that comment
can be made on what is occurring from a phenomenological perspective. More data is required in order
to come to more definitive conclusions in this regard.

Future Work

This study will be expanded to include data for 9No. beams, each with different straight-profiled
prestress strand eccentricities. The relationship between prestress force eccentricity and natural
frequency will therefore be investigated. The existence of interaction effects between prestress force
magnitude and eccentricity and their effect on natural bending frequencies will also be investigated.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the financial support provided by the Irish Research Council (IRC)
under their Embark initiative. The authors would also like to sincerely thank Banagher Concrete,
Heitons Steel, Roadstone, Fairyhouse Steel and Freyssinet for their support in supplying testing
materials throughout the duration of the project to date.

REFERENCES

Hamed, E., & Frostig, Y. (2006) “Natural frequencies of bonded and unbonded prestressed beams -
prestress force effects”, Journal of Sound and Vibration, Vol. 295, No. 1-2, pp. 28-39.
Hop, T. (1991) “The effect of degree of prestressing and age of concrete beams on frequency and
damping of their free vibration”, Materials and Structures, Vol.24, No. 3, pp. 210-220.
Kim, Jeong-Tae, Yun, Chung-Bang, Ryu, Yeon-Sun and Cho, Hyun-Man. (2004) “Identification of
prestress-loss in PSC beams using modal information”, Structural Engineering and Mechanics,
Vol. 17, No.3-4, pp. 467-482.
Miyamoto, A., Tei, K., Nakamura, H. and Bull, J. (2000) “Behavior of Prestressed Beam Strengthened
with External Tendons”, Journal of Structural Engineering, Vol. 126, No. 9, pp. 1033-1044.
Quilligan, A., O'Connor, A., & Pakrashi, V. (2012) “Fragility analysis of steel and concrete wind
turbine towers”, Engineering Structures,Vol. 36, No. 0, pp. 270-282.
Saiidi, M., Douglas, B. and Feng, S. (1994) “Prestress Force Effect on Vibration Frequency of
Concrete Bridges”, Journal of Structural Engineering, Vol. 120, No. 7, pp. 2233-2241.
Tse, F.S., Morse, I.E. and Hinkle, R.T. (1978) Mechanical vibrations: theory and applications, Allyn
and Bacon.

ACMSM23 2014 338